Acceleration of energetic particles by whistler waves in active space experiment with charged particle beams injection

Available online at www.sciencedirect.com

www.elsevier.com/locate/asr

Advances in Space Research 49 (2012) 859–871

Acceleration of energetic particles by whistler waves in activespace experiment with charged particle beams injection

Nikolai Baranets a,⇑, Yuri Ruzhin a, Nikolai Erokhin b, Valeri Afonin b, Jaroslav Vojta c,Jan Smilauer c, Karel Kudela d, Jan Matisin d, Mircea Ciobanu e

a Institute of Terrestrial Magnetism, Ionosphere, and Radio Wave Propagation, Russian Academy of Sciences (IZMIRAN),

Troitsk 142190, Moscow region, Russiab Space Research Institute, Russian Academy of Sciences, Moscow 117997, Russia

c Institute of Atmospheric Physics, Academy of Sciences of the Czech Republic, 14131 Prague, Czech Republicd Institute of Experimental Physics, Slovak Academy of Sciences, 04001 Kosice, Slovakia

e Institute of Gravitation and Space Sciences, 71111 Bucharest, Romania

Received 10 September 2010; received in revised form 8 October 2011; accepted 1 December 2011Available online 9 December 2011

Abstract

In this paper the investigation of wave-particle interaction during simultaneous injection of electron and xenon ion beams from thesatellite Intercosmos-25 (IK-25) carried out using the data of the double satellite system with subsatellite Magion-3 (APEX). Results ofactive space experiment devoted to the beam-plasma instability are partially presented in the paper Baranets et al. (2007). A specific fea-ture of the experiment carried out in orbits 201, 202 was that charged particle flows were injected in the same direction along the magneticfield lines B0 so the oblique beam-into-beam injection have been produced. Results of the beam-plasma interaction for this configurationwere registered by scientific instruments mounted on the station IK-25 and Magion-3 subsatellite. Main attention is paid to study theelectromagnetic and longitudinal waves excitation in different frequency ranges and the energetic electron fluxes disturbed due towave-particle interaction with whistler waves. The whistler wave excitation on the 1st electron cyclotron harmonic via normal Dopplereffect during electron beam injection in ionospheric plasma are considered.Crown copyright � 2011 Published by Elsevier Ltd. on behalf of COSPAR. All rights reserved.

Keywords: Electron beam injection; Wave-particle interaction; Whistler waves

1. Introduction

Different aspects of a beam-plasma instability (BPI) forelectron beam injection through the extended hollow beamof xenon ions were considered in the paper Baranets et al.(2007) for the case of orbit 202. One of the most interestingresult was related to the absorption or excitation of high-frequency (HF) waves under the electron–cyclotron reso-

0273-1177/$36.00 Crown copyright � 2011 Published by Elsevier Ltd. on beh

doi:10.1016/j.asr.2011.12.001

⇑ Corresponding author. Tel.: +7 495 3340913; fax: +7 495 3340124.E-mail addresses: [email protected] (N. Baranets), ruzhin@

izmiranu.ru (Yu. Ruzhin), [email protected] (N. Erokhin), [email protected] (V. Afonin), [email protected] (J. Vojta), [email protected](J. Smilauer), [email protected] (K. Smilauer), [email protected](K. Kudela), [email protected] (J. Matisin), [email protected](M. Ciobanu).

nance condition in dependence of the ratio of Larmorradius of the beam electrons to the lateral wave length(with respect to magnetic field B0) in accordance with the-oretical results by Kitsenko and Stepanov (1961). This fea-ture of BPI mechanism is confirmed by the measurementsof HF electric fields ehf and fast electron and ion fluxesJe,p carried out at Magion-3 subsatellite. Preliminary resultsof electromagnetic wave instability (EMI) development rel-ative to a whistler wave excitation and some experimentaldata were considered in Baranets et al. (2009). The mainresults of this investigation are the satisfactory amplitudecorrelation between registered on IK-25 the low-frequency(LF) waves, strong VLF wave excitation and the growthrates of whistler waves calculated for a quasi-hydrody-namic model of beam-plasma interaction. It is suggested

alf of COSPAR. All rights reserved.

860 N. Baranets et al. / Advances in Space Research 49 (2012) 859–871

that the xenon ion beam injection leads to very low-frequency wave excitation with the maximum energy spec-trum to be rather far away from the electron beam-inducedpeak of the energy wave spectrum, such that there is noparametric waves coupling in the linear approximation.Therefore, the effect of the ion beam injection on the HFwave spectra can be conventionally neglected.

In these papers the result of active space experimenthave been obtained for a temporal behavior of registereddata and as a function of the parameters under sometheoretical consideration. The main attention in this articleis paid to the study the beam-induced electromagneticinstability relative to a whistler wave excitation for thehigh-altitude ionospheric conditions. A wide range ofexperimental data obtained from both satellites allowedto carry out the complex investigation of these effects dur-ing electron beam injection. To study the resonant andnonresonant effects of the wave-particle interaction the dif-ferential charged particle fluxes recorded on Magion-3 arethe subject of more detailed investigation in this paper.One of the most interesting result is related to the anoma-lous amplification of ULF–VLF electromagnetic fields andquasi-stationary electric/magnetic fields observed in thefrequency range 0.8 [ jxm � xbej/xLH [ 1.2, where xm,xbe, xLH are the frequencies of electron beam modulation,slow space-charge beam waves, and lower hybrid reso-nance, respectively. Although it is not quite clear whichtype of waves excited in the outer space is associated withthe frequency of beam modulation in the electron gun,one can assume that the nonlinear wave coupling withthe frequencies xm, xbe can lead to beat waves close tothe low-hybrid frequency xLH. In the high-frequency rangethe excitation of whistler waves occurs on the 1st electroncyclotron harmonic. Excitation of backward-propagatingwaves via normal Doppler effect and subsequent whistlerwaves absorption/scattering by the fast charged particlesdue to Cerenkov resonance are confirmed by the data offlux particles and fields registered on Magion-3. Thesetwo resonance mechanisms of wave-particle interactionsgive the answer to some questions stated in Prech et al.(2002) concerning to the anomalous electron fluxesdetected on Magion-3 subsatellite during APEX experi-ment on orbits 431, 514.

Many parameters of electron injection in our experimentare close to those of the Polar-5 rocket probe experimentcarried out according to the ‘mother–daughter’ scientificpayload scheme (Maehlum et al., 1980). During 8–10 keVelectron beam injection produced on the ‘daughter’ pay-load the amplification of low-energy electron fluxes([1 keV) were registered on ‘mother’ payload at the dis-tance 10–15 m to the magnetic field line guiding the injectedelectrons. It was suggested that at high altitudes the beam-induced longitudinal waves can cause the wave-particle col-lective effects which give a rise of thermal plasma electronfluxes. A similar effects of the anomalous amplification ofthermal plasma ions have been observed during APEXexperiment on the data 202 orbit. But these resonance

phenomena are related to the whistler waves excited dueto the electron beam injection and considered in this paper.Electron beam experiments have been conducted at rock-ets, satellites, and space shuttle orbiter to study the prob-lem of exchanging large currents between emittingsurfaces and the environment (Grandal, 1982; Mishin etal., 1989). In considered here experiments carried out athigh altitudes �1450–1800 (201, 202 orbits), the strongbody potential changes at IK-25 was not obtained maybe due to the Xe-neutral gas release before the electroninjection is switched on. It is believed that the neutral gasrelease or injection of a plasma can effectively neutralizea charged particle beam (Marshall et al., 1988).

2. Scientific equipment and spatial configuration of theinjections

The ion injector was a Hall-type stationary plasmathruster (SPT) with a longitudinal acceleration of xenonions. The injection current varied in the range Ibi � 2.0–2.6 A, and the output ion energy reached 250 eV. A consid-erable amount of electrons with the density ne and temper-ature �3–5 eV have been injected from SPT channel tocompensate the ion beam charge. The electron injector isa straight-channel three-electrode electron gun (EG) oper-ating at modulation frequencies varying from 32 Hz to250 kHz (over a time period of 2–12 s) after the first secondof dc injection, with 1-s intervals between the injectioncycles. The basic cycle of electron injection is 23 s. TheEG control electrode provided 100% modulation of theelectron beam current, Ibe � 100 mA, thereby forming sep-arate injection micropulses with a duration of 2 ls. Elec-tron beam in all cases reported in the present study wasinjected in the direction opposite to the satellite velocity.The charged particle beams have been injected during aXenon neutral gas release started �35 s before; the flowrate of gas release is 3 mg/s. A DOK-S charged particlespectrometer installed at the subsatellite measured the dif-ferential fluxes of fast electrons and protons in eight energyintervals in the ranges 25–420 and 20–1300 keV, respec-tively. To record the fast particles, two pairs of detectors1E, 1P and 2E, 2P installed at Magion-3 were oriented intwo mutually perpendicular directions. Detectors 1e, 1pare oriented along the main axis of the subsatellite body.The magnetic field components at the station and the sub-satellite were measured by SGR-5 and SGR-6 ferroprobemagnetometers with accuracies of 1 nT and 2/16 nT,respectively. The measurement accuracy of the magnetom-eter installed at the subsatellite changed automaticallydepending on the actual value of the magnetic field jB0j.The electron temperature Te and the ion density of the ther-mal plasma ni(V) (where 0 < V < 12 V is the sweep voltageat the grids of the ion trap) were measured using the KM-10 and KM-13 complexes installed at the station and thesubsatellite, respectively. Different ionospheric parameterssuch as a quasi-steady electric field E (0.1–2 Hz) and elec-tric/magnetic component of ULF–VLF waves elf/blf used

Table 1Main parameters of the beam-plasma system in the ionosphere.

Parameter Value

Unperturbed plasma density (cm�3) n0 ’ nix (V 6 1 V) (3–7) � 103

Electron temperature components in the unperturbed plasma (K) Tex, Tey, Tez (5,3.4,4.5) � 104

Injection currents of xenon ions and electrons (A) Ibi, Ibe 2.1–2.4, 0.1Maximum ion/electron accelerating voltage (kV) Ui, Ue 0.25, 10Alfven velocity and flow velocities of electrons and xenon ions (m/s) vA, u, viz 7.5 � 106, 4.2 � 107,

1.1 � 104

Average transverse dimensions of the ion and electron beams (m) rcx, rce 270, 10Gyrofrequency of xenon ions (rad � s�1) xcx 14–20Plasma frequency of the beam ions (rad � s�1 ) xpx 13–16Gyrofrequency of the hydrogen plasma component (rad � s�1) xci (1.9–2.5) � 103

Plasma frequency of the hydrogen plasma component (rad � s�1) xpi (0.13–0.16) � 106

Electron gyrofrequency (rad � s�1) xce (3.4–4.4) � 106

Electron plasma frequency of the unperturbed plasma (rad � s�1) xpe (5.6–6.5) � 106

Plasma frequency of the beam electrons in the dc injection mode for hollow beam model (rad � s�1) xbe (0.13–0.14) � 106

Modulation frequency of the electron current during 2–12 s of ac-injection (after 1 s of dc-injection)(Hz)

xm/2p 32–250,000

N. Baranets et al. / Advances in Space Research 49 (2012) 859–871 861

in this paper are measured by the scientific instrumentsmounted at both satellites. A more detailed characteristicsof the scientific payload of the mother–daughter satellitesystem are described, for example, in Oraevskiy andRuzhin (1992) and Prech et al. (2002).

Spatial field components measured in the orthogonalcoordinate systems X, Y, Z and x0, y0, z0 at the satelliteand subsatellite are reduced to the new ones for the left-handed Cartesian coordinate system x, y, z with zkB0.The angles b3, b03 between the field B0 and axes Z and z0

of the satellite/subsatellite coordinate systems were deter-mined by the SGR-5,6 magnetometers, respectively. Themain characteristics of the injectors and other parametersof the beam-plasma system (obtained for orbit 202) aregiven in Table 1. The distance between IK-25 andMagion-3 satellites is varied near 110–120 km for consid-ered orbits. Unfortunately, the data of the subsatellite posi-tion on orbits cannot be presented in this paper. However,it is known that the subsatellite orbit is strongly elliptically-extended along the main satellite orbit. The main satellitewas stabilized on 3 axes in space with Z-axis oriented alongthe gradient of the magnetic field B0, while the subsatellitewas stabilized on z0-axis oriented parallel to the magneticfield.

Fig. 1. Electron beam injection (e�) directed through the hollow beam ofxenon ions (Xe+) with the beams density and velocity nbe, v and nbi, vi,respectively. Release of the electron flow with the density and streamvelocity ne, ue for a compensation of the ion beam charge at the output ofSPT is shown by the up arrow. Here, B0 and vs are the directions of thequasi-steady magnetic field and the satellite velocity in the X, Y, Z

coordinate system, Z-axis points away from the Earth.

3. Main features of the beam-plasma interaction

The system of an electron beam nested in an ion beam isaxially asymmetric with respect to the magnetic field direc-tion due to the small velocity of xenon ions (viz/u � 3 � 10�4), which is comparable to the velocity of thesatellite (viz/vs � 1.5) moving at an angle to the magneticfield (Fig. 1). For certain injection parameters, the gener-ated HF oscillations can reach a saturation level (corre-sponding to the onset of a nonlinear regime), after whichthe spectrum begins to extend toward lower frequencies.In this case, the hollow beam of heavy xenon ions injectedat pitch angles of up to Dapi ’ 60� with a maximum flux

density within the angles Dapi 6 30� will play the role ofa damping layer for waves induced by the electron beamin the entire interaction region in the vicinity of the satel-lite. In this regard, the generation of extremely low-frequency (ELF) waves and the possibility of controllingthe nonlinear interaction mechanism are of most interest.

862 N. Baranets et al. / Advances in Space Research 49 (2012) 859–871

3.1. Beam current profiles

When a low-energy electron beam (�10 keV) is injectedinto the ionospheric plasma, the development of instabili-ties and the excitation of electromagnetic fields depend sub-stantially on the shape and density of the beam. On theother hand, in a complex current system, the current profiledepends on the energy density of the excited waves, whichin turn modulate the electron beam, thereby producing afeedback in the beam-plasma system. In order to determinethe main characteristics of wave excitation and chargemodulation, we assume that the electrons are injected inthe presence of electric fields (excited over �1 s), whichmodulate the beam in the injection region. After severalgyrations of dense particle beams, electrostatic repulsionforces transform them into hollow flows with the averagedensity nba (with a = e, i), determined by the expression

Iba ffi 2pR r2

r1evzðrÞnbaðrÞrdr, where r1 and r2 are the mini-

mum and maximum radii of charged particle gyration atthe internal and external boundaries of the beam,

r ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffix2 þ y2

p. The averaged flow velocity was determined

as u � hvzia ¼ ð1=Da0ÞR

Da0 v cosðap þ aÞda, where the effec-

tive range of pitch angles is Da0 > Da0 (�2�–3� and �60�for electrons and ions, respectively, at z = 0). The effectiverange of pitch angles for electrons,

Da0 ’ Da0 þpe

mvxcedEx þ vdBz sin ape � vdBy cos ape

� �; ð1Þ

reflects the amplitude modulation of the beam by quasi-steady fields in the injection region over a time t � 1 s,where e, m are the charge and mass of electron, and xce

is the electron gyrofrequency. The upper bar, dFx;y;z ¼Fx;y;z � Fx;y;z, denotes the time-averaged quantity over theinterval Dt � (7–10) � Dt0 for the any field measurementsF , where Dt0 is the duration of one telemetry frame atthe IK-25 and v is the electron velocity at z = 0 (at the out-put of the EG modulator). The smoothing of the physicalvariables in formula (1) is of great importance for eliminat-ing the influence of fast fluctuations on the calculateddensity of the injected beam; without such a smoothing,the possibility of numerical instability can arise. Thefluctuation amplitudes of the transverse and longitudinalvelocities of the injected particles in an equilibriumstate and a flow thermal electron velocity were estimatedusing the expressions dv\,z 6 max{v\,z � hv\,zia}, and

vbe � dv2z þ dv2

?� �1=2

, respectively, here v? � v2x þ v2

y

� �1=2

.

These allow one to use the effective angular divergenceDw � cos3(ape � Da0/2) � cos3(ape + Da0/2) as a certainmeasure of beam heating.

3.2. Electron beam current modulation

An inhomogeneous electron beam in the EG channel isproduced in two stages. During short-term interaction withthe driving field of the modulator sixm 1 (si = 2 ls),

low-energy electrons are first modulated over velocity;then, the beam is additionally accelerated in the spacebetween the anode and the grid modulator. The mostimportant parameters of the electron-field interaction inEG modulator are the modulation depth R, the change inthe field phase H1, and the efficiency with which the drivingfield acts on the beam electrons G ¼ 2 sinðH1=2Þ=H1

(Kaptsov, 1992). In the regime of EG-modulation for thesimplest case of free electron gyration in the outer space,the spectral composition of the electron current Ibe isdetermined by the expression

Ibeðz; tÞ ¼ I0 þ 2I0

X1n¼1

ð�1ÞnJ nðnX DðzÞÞ

� cos n xmt �HDðzÞ �H1

2

� �; ð2Þ

where Jn(nXD(z)) is the Bessel function, and XD(z) =RGHD(z)/2 is the bunching parameter over the length z.The current I0 is defined during the 1-s of dc-injection.The amplitudes of the harmonics of the convection currentare determined by the expression In

beðz; tÞ ¼ 2I0J nðnX DðzÞÞ.In the regime of modulation, the plasma frequency of thebeam electrons xbe is determined for the first harmonicof the convection current, I1

be. However, the beam particlesdynamics in the ionospheric plasma strongly depends onthe excited waves due to a plasma electromagnetic instabil-ity and self-fields associated with the beam charge and cur-rent densities, (Banks and Raitt, 1989). The beam-inducedwaves propagating back to the injector can lead to the elec-tron velocity modulation along beam body thereby pro-duced a feedback in the beam-plasma system. Despite theelectron velocity divergence in term of Eq. (1) is used toevaluate the effective pitch-angles, the self-consistentbeam-plasma interaction problems are out of the scope inthis paper. The current Ibe(z, t) in a modulation mode isof first-order approximation. A beam charge compensationby the ionospheric plasma ions is a complicated problemdo not considered here.

3.3. Beam-driven electromagnetic instability

Electron beam moving through ionospheric plasmaalong magnetic field lines can excite electromagnetic wavesin the frequency range xci x xce for the case of rela-tively weak magnetic field (xce xpe). It is known that in ahydrodynamic approach the transverse waves with k\ = 0do not excited by the cold beam in which all electrons havea stream velocity u. However, the electromagnetic instabil-ities are very sensitive to an anisotropic particle velocitydistribution. In this case the instability can be excited evenfor cosh ’ 1. Dielectric tensor components in the presenceof beam electrons contain a small correction �ij ¼ �0

ij þ �0ij,where �0

ij and �0ij are the tensor components for coldplasma and beam particles, respectively. In the absenceof beam electrons the reduced form of the commondispersion equation relative to whistler wave excitation

N. Baranets et al. / Advances in Space Research 49 (2012) 859–871 863

x2 x2pe; h – 0

� �can be presented in the form (Mikhai-

lovskiy, 1975)

N 2 cos2 h ffi �011 i�0

12; ð3Þ

where N2 = c2k2/x2, and the plasma dielectric componentsare

�011 ¼ �x2

pe= x2 � x2ce

� �;

�012 ¼ �ix2

pexce=x x2 � x2ce

� �:

For sufficiently dense plasma ðx2pe � x2

ceÞ and quasi-longi-tudinal wave propagation, the solution of Eq. (3) for whis-tler waves can be presented by the relation

xðk; hÞ ¼ jxcejc2k2 cos h

x2pe þ c2k2

: ð4Þ

The corresponding equation for whistler waves excitationin the beam-plasma system is the following

N 2 cos h� i�012 � �0�= cos h ¼ 0; ð5Þ

where �0� ¼ 12�011 2i cos h�012 þ cos2 h�022

� �are the correc-

tions of dielectric tensor components due to the presenceof beam electrons. The warm beam components derivedin the kinetic approach are

�011 ¼4pe2

mx

X1n¼�1

fnU?v2?

n2

n2J 2

nðnÞ* +

;

�012 ¼4pe2

mx

X1n¼�1

fnU?iv2?

nJ nJ 0nn

* +;

�022 ¼4pe2

mx

X1n¼�1

fnU?v2?J 02n ðnÞ

* +;

ð6Þ

where fn = (x � nxce � kzvz)�1, n = k\v\/xce and

h. . .i ¼Rð. . .Þv?dv?dvz. Here kz is the longitudinal compo-

nent of the wave vector for waves propagating at an angleh to the magnetic field in the Cartesian coordinate system(x,y,z), and

U? ¼1

v?

@F 0

@v?þ kz

x@F 0

@vz� vz

v?

@F 0

@v?

: ð7Þ

The beam distribution function is defined as F 0 ¼ n0bedðvz

�uÞf?ðe?Þ; e? ¼ v2?=2, such that

v2?=2 �

Ze?f?ðe?Þde?–0;

IA ¼Z

e?f 0?ðe?Þde?–� 1;

ð8Þ

where n0be, and f? v2

x þ v2y

� �are the unperturbed electron

beam density and two-dimensional beam distributionfunction, respectively. Using the components �0ij defined

in Eq. (6) the expression for �0� can be derived in the form

�0� ¼4pe2

mx

X1n¼�1

fnU? ðn=nÞJ nðnÞ � cos hJ 0nðnÞ� �2 � v

2?2

* +:

ð9Þ

After summing up for n = 0,±1 and integrating over veloc-ity phase space h. . .i, this expression can be reduced to thefollowing (h – 0)

�0� ¼ �x2

be

4x2

IAðx� kzuÞð1� cos hÞ2

x� xce � kzuþ

k2z v2

?=2� �

ð1� cos hÞ2

x� xce � kzuð Þ2

24þ IAðx� kzuÞð1 cos hÞ2

xþ xce � kzuþ

k2z v2

?=2� �

ð1 cos hÞ2

xþ xce � kzuð Þ2

35;ð10Þ

which for the case of h = 0 turn to �0þ described in Mikhai-lovskiy (1975), x2

be ¼4pe2no

bem . Third and fourth terms of the

right side in Eq. (10) can be ignored if the first two are res-onant and vice versa. As a result the dispersion relation forthe whistler mode wave excitation can be presented in theform ðj�0

11j j�012jÞ:

c2k2 cos hx2

þx2

pexce

x x2 � x2ce

� �ffi �x2

bev�4x2

IAðx� kzuÞx� xce � kzu

þk2

z v2?=2

� �x� xce � kzuð Þ2

24 35; ð11Þ

where v± = (1 ± cosh)2. It is evident from Eq. (11) that theelectromagnetic wave excitation occurs in the frequencyrange x � xce � kzu � 0 for the case of �0þ (normal Dopp-ler effect, n = 1), i.e. for kz � �xce/u. It is follows that inthe range of whistler mode, the backward-propagatingwaves are excited due to the electromagnetic instability(well-known also as a beam-anisotropic instability).

For a weak-beam approach nbe n0 and an absolutecharacter of the instability, the solution of the Eq. (11)can be obtained for the frequencies x � kzu + njxcej + e,where the frequency x corresponds to the solution of dis-persion relation for whistler waves in ionospheric plasma,and e = dx + ic (jej jxj, n = 0,±1). For moderatedetunings, when je/(x � kzu)j 1, or jeðx� kzuÞj jðx� kzuÞ2 � x2

cej, by linearizing (11) with respect to thesmall parameter jgj = je/xj 1 and neglecting the termson the order of �g3, g4, the dispersion equation can bereduced to the form Ag2 + Bg + C = 0, where A, B, Care fairly complicated functions of the parameters xpe,xce, x, xbe, u, v\. In this case, the normalized growth ratec/x � Im g of excited waves is equal to

Img ¼ þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffij B2 � 4AC j

p2A

; B26 4AC: ð12Þ

-150-100-50

050

100150

-20

0

20

40

δBx’

, nT

δBx,

nT

δBx’δBx

-150-100-50

050

100

-250255075

δBz’

, nT

δBz,

nT

δBz’δBz

105

104

103

102

101

J e,(

cm2 ⋅k

eV⋅s⋅

sr)-1

1e

2

3

4

5

j ix’,

10-5

A/m

2

0

2

4

6

1600 1650 1700 1750 1800

I b, (

tel)

V

t, s

IbeIbiXe

864 N. Baranets et al. / Advances in Space Research 49 (2012) 859–871

Despite of the dispersion relation (11) was obtained forkinetic approach, the solution of this equation relative tothe growth rate has a quasi-hydrodynamic character c � ch,i.e. the major part of beam electrons are responsible forinstability development.

4. Experimental results

During electron beam injection through the xenon ionbeam the perturbation of quasi-steady magnetic field com-ponents have been observed both near the injection pointat IK-25 satellite and at far distances near the Magion-3

subsatellite. In Figs. 2 and 3 for orbits 201, 202, respec-tively, the quasi-steady magnetic field disturbances regis-tered by the ferroprobe magnetometers in two points ofionospheric plasma are presented. Satellite-registered mag-netic field disturbances dBx;y;z ¼ Bx;y;z � Bx;y;z in consideredtime do not exceed 20–30 nT (z � 0), while the subsatel-lite-observed magnetic field fluctuations are much moreintensive and achieve hundreds nT in magnitude. Resultsof the magnetic field disturbances have been presented ear-lier in Oraevskiy et al. (2001) for the case when the electronand xenon ion flows were injected in opposite directions

-150-100

-500

50100

-150-5050150

δBy’

, nT

δBz’

, nT

δBy’

δBz’

104

102

100

J p,(

cm2 ⋅k

eV⋅s⋅

sr)-1

1p

0.4

0.5

0.6

0.7

-10

0

10

b lf,1

0-4nT

/Hz1/

2

Ps,V

Ps

0.4

0.5

0.6

0.7

-10

0

10

b lf,1

0-4nT

/Hz1/

2

Ps,V

4

6

8

4

6

8

j ix’,

10-5

A/m

2

0

2

4

6

8

1620 1630 1640 1650 1660 1670 1680 1690

I b, (

tel)

V

t, s

IbeIbiXe

Fig. 2. Dependencies of the lateral (poloidal) and longitudinal quasi-steadymagnetic field components (0.1–2 Hz) measured at Magion-3 satellitesdBy0 ;z0 , differential ion fluxes Jp with the energy e1p = 340 keV and the densityof thermal plasma ion fluxes jix0 measured for direction �x0 are presentedversus a time of active space experiment duration, orbit 201. Active mode ofelectron gun operation (current Ibe), ion beam injection (Ibe) and xenon gasrelease (Xe) in telemetry volts are presented in the lower part. Time durationt is counted relative the instant 1 h 26 min 59 s (UT) before injections areswitched on, the orbit altitude is H = 1450–1715 km, distance betweensatellite and subsatellite is d ’ 110 km, b3 � b3ðdB0; ZÞ ’ 140� 155 ,b03 � b3

dB0; z0� �

’ 140� 165 , McIlvain parameter is L = 1.39–1.51.

Fig. 3. Dependencies of the lateral and longitudinal quasi-steady magneticfield components measured at Magion-3 and IK-25 satellites dBx0 ;z0 ; dBx;z,respectively, differential electron fluxes Je (�1e = 245 keV) and density of thethermal plasma ion fluxes jix0 as well as the injectors job active cycle arepresented versus time at orbit 202. Time t is counted relative the instant 3 h24 min 22 s (UT) (before injections), H = 1490–1798 km, distance betweensatellites is d ’ 120 km, b3 � 140 –165 ; b03 � 135 –170 ; L ¼ 1:32–1:46.

along the magnetic field and its magnitudes achieved�500 nT in the satellite region (orbit 266) what is by twoorder of magnitude higher then the nominal valuesdB � Ibe/rce ’ 10–20 nT. Anomalously large values wereregistered for the case of injection when the electron fluidvelocity was close to Alfven velocity u � vA. The excitationof ELF magnetic field component in the range x 6 xci

observed on orbit 266 may be caused also by the Alfvenresonance. It was noted that the wave excitation stronglydepends on the beam thermal spread, vbe. If the beam elec-trons are colder, the ELF wave excitation level is higher.This experimental result is in accordance with the kinetictheory of beam-plasma interaction. Magnetic field pertur-bations can be considered as a superposition of Alfvenwave packet envelope and slowly-varying fields inducedby nonlinear plasma currents (Mett and Tataronis, 1989;Moffatt, 1978). All presented data prove the existence ofconsiderable disturbances of the magnetic field componentsinduced by the electron injection and registered at Magion-

3, as well as more weak ones at IK-25 caused by the xenonion injection. These disturbances correlate well enoughwith the injection cycles presented on lower plot of Figs.2,3. The blank data in Fig. 3 are caused by the particularcircumstances in a subsatellite telemetry data transmission.

The upper part in these figures demonstrates also thedifferential fluxes of fast ions and electrons detected for

N. Baranets et al. / Advances in Space Research 49 (2012) 859–871 865

direction ‘1p’, ‘1e’ along the subsatellite z0-axis. In the mid-dle part slow amplitude modulations have been presentedfor the VLF magnetic field component blf (x/2p =9.6 kHz) and satellite body potential Ps registered at IK-

25. Subsatellite measurements have been presented alsoby the thermal plasma ion flows jix0 registered by the iontrap (KM-13) in �x0 direction. It is worth to note the fea-tures of the thermal plasma ion flows shown at these fig-ures. The data registered on orbit 201 demonstrate thesynchronous behavior of jix0 with the periodic variationsof xenon ion current Ibi. Similar slow amplitude modula-tions were observed for the VLF magnetic field componentand satellite body potential Ps registered at IK-25 for thisorbit despite the xenon ion beam was injected in direct cur-rent mode. The nature of these variations is related to theion acceleration characteristics in SPT channel; slow syn-chronous variations of thermal plasma ion flows atMagion-3 can be modulated as by the wave packet stimu-lated due to electron injection as by the ion beam-inducedVLF waves. Another behavior of thermal plasma ion flowshas been recorded on orbit 202 (Fig. 3), i.e. the resonantcharacter of flow changes is observed when the modulationfrequency of electron beam current sweeps to xm/2p = 62.5 kHz. This modulation frequency is close to thefrequency values �xce/2p/6 � 90–100 kHz used as one ofthe whistler excited wave frequencies during their growthrate calculation. This increase of thermal plasma ion flowsis similar to that observed in Polar-5 rocket experiment onamplification of low-energy electron flows ([1 keV). Tostudy these resonance phenomena and to answer on thequestion what can be considered as a strong effect, it isneeded to carry out a separate investigation.

In the article Baranets et al. (2009) the data of slow Ps-variations versus ion beam injection as well as more nega-tively biased ones during a dc-injection of electrons werepresented for another time of the experiment on orbit201. Despite the data of the potential Ps shown in Fig. 2are not so regular, it is evident that the satellite body poten-tial varies in accordance with the periodic ion beam currentmodulations. Changes of the potential Ps during electroninjection is similar to that of the negative dc bias of an elec-trode which occurs during the injection of HF power in aplasma. The induced electromagnetic field, once generated,is sustained by the electron beam injection, thereby produc-ing a nonlinear pressure on the ambient plasma. Appear-ance of HF fields in the vicinity of injector can lead to anamplification (or suppression) of low frequency waves.The behavior of the satellite body potential at IK-25

strongly differs from that in rocket-borne experiments withcharged particle beams due to the effects of Xenon quasi-neutral plasma injection, and gas release ensuring the neu-tralization processes, at least, for considered conditions.

As concerning the fast charged particles detected by fourspectral analyzer at Magion-3, the indirect evidence of elec-tron precipitations with energies one order of magnitudehigher then the energy of injected electrons (0.5 A,27 keV) was previously observed in Araks rocket experi-

ment. Later, X-ray bursts observed in Trigger experimentwere also attributed to the precipitation of fast particles(Grandal, 1982; Roeder et al., 1980). Registration resultsof fast charged particles (electrons) during APEX-experi-ment for a wide range of ionospheric conditions wasreported early in Prech et al. (2002). It is worth to mindthat all these experiments including the ones at IK-25 andMagion-3 satellites (APEX project) have been conductedfor different ionospheric conditions. So, the comparisonof these results may be done only in the first approach.

It is obvious that the disturbances dBx;y;z; J e;p; jix0 arecaused by the electron/ion beam injection, but to obtainmore detailed characteristics of the beam-plasma interac-tion in remote regions of ionospheric plasma the hybridcomputational algorithm was used at the second stage ofcomplex data processing. The measured and calculatedparameters during injection on, hj and sj respectively, weretransformed (with allowance for simulations of activespace experiment) into a new sequence, hj(t), sj(t), p(t))hj(p), sj(p), in ascending order of the parameter p. As resultof more complex data processing, all following figuresdemonstrate new dependencies as a function of the calcu-lated parameter p. To study the different characteristicsor experimental data by this manner, first of all, the stableionospheric conditions are needed during a consideredtime. Second, the satellite/subsatellite telemetry data havebeen transmitted with the different sampling rates. There-fore, in order to form a common digital stream from bothsatellites, the transmitted data (measurements) were previ-ously synchronized.

In the top panels of Figs. 4,5 the growth rate calculatedvalues ch, ck for electromagnetic (whistler mode) and beampotential waves are presented versus a triplet wave ratiop � p1 = jxm � xbej/xLH, where xLH is a low-hybrid fre-quency. Excitation of whistler waves occurs for ionosphericconditions considered in Section 3.3. when the backwardpropagating waves are excited due to the normal Dopplereffect and directed to the source of injection. To obtain nor-malized growth rates presented at these figures Eq. (11) issolved for small detunings by means of 500 � 500 iterationsover x and k near their resonant values at every time step.For comparison of the electromagnetic instability develop-ment with beam-plasma instability, the excitation of longi-tudinal waves was considered on the basis of kineticapproach by Kitsenko and Stepanov (1961). For a ‘coldplasma-warm beam’ system the growth rate of these wavesnear the frequencies x = x± + e can be expressed by thefollowing

ck ¼�ffiffiffipp

x2bex�

2k2v2be

xjnjbe

2jnjjnj!x2

pe

x2�

cos2 hþx2

pex2� sin2 h

x2� � x2

ce

� �2

" #�1

� z0 exp �z2n

� �; ð13Þ

where

zn ¼x� � njxcej � kzuffiffiffi

2p

kzvbe

;

1.0

1.5

2.0

γh /ω, 1

0-2

2.0

4.0

6.0

γk /ω-,

10-2

0.4

0.5

0.6

b lf,1

0-4nT

/Hz1/

2

ω/2π=9.6 kHz

0.5

0.6

0.7

0.8

e lf,1

0-2m

V/m

/Hz1/

2

149 Hz

0.5

1.0

1.5

e lf,1

0-2

75 Hz

104

102

100

10-2

0 1 2 3 4

J 1e,

(cm

2 ⋅keV

⋅s ⋅sr

)-1

|ωm-ωbe|/ωLH

ε1e =245 keV

Fig. 4. Normalised growth rates of the electromagnetic instability ch/x(x ’ xce/6,n = 1,h ’ 0.04) and beam-plasma instability ck/x� (x ’x�,n = 0,h ’ 0.36) at the top panels; amplitudes of the magnetic andelectric component of ULF–VLF waves blf,elf measured at IK-25, as wellas the anomalous electron fluxes measured at Magion-3 (‘1e’-ap1 ’20�–30�) are presented versus p1 = jxm � xbej/xLH, orbit 201. Boldhorizontal segments in bottom and top panels denote the range(p1 ’ 0.8–1.2) where the resonance effects are observed.

6.0

8.0

10.0

γh /ω, 1

0-2

10-1

10-2

γk /ω-

-10

0

10

20

δB⊥

, nT

-60

-40

-20

0

20

δEz,

mV

/m

104

102

100

10-2

J p,e

2 ⋅keV

⋅s⋅sr

)-1

ε1e = 245 keV

106

104

102

100

10-2

0 1 2 3 4|ωm-ωbe|/ωLH

ε1p = 20 keV

Fig. 5. Normalised growth rates of EMI and BPI instabilities calculatedfor constant resonance detunings (jdxj � dx0) relative to the whistler andlongitudinal wave excitation (h ’ 0.34); perturbations of the lateral andlongitudinal quasi-steady magnetic, electric field components dB\,dEz,respectively, measured at IK-25, and the anomalous fluxes of electrons andions (‘1e, 1p’) are presented versus the parameter p1 = jxm � xbej/xLH,orbit 202. Bold horizontal segments (p1 ’ 0.8–1.2) denote the same as atFig. 4.

866 N. Baranets et al. / Advances in Space Research 49 (2012) 859–871

xbe = (kxvbe/xce)2, and x±(h) correspond to the higher (+)

and lower (�) hybrid plasma resonances.It was possible to determine the growth rate ch by means

of the iterations in certain range of frequency detunings dxwhen B2

6 4AC. Thus the growth rate calculation wascarried out in two ways: (1) as soon as dx(dk) correspondsto the condition of Eq. (12) (presented in Fig. 4), and (2)for some constant frequency detuning dx � dx0 (Fig. 5).As a consequence, the solution of Eq. (12) for ch as wellas the calculation ck by the first method are very sensitive

to small variations of the ionospheric parameters the dataon which are supplied in real-time mode to the input ofthe numerical algorithm to define the parameters of insta-bilities. It was assumed in Section 3.3. (first of all, for sim-plicity) that the development of transverse-wave instabilityhas an absolute character during the oblique electron beaminjection relative to the xenon ion stream, that is the whis-tler wave amplitudes grow in time wherever inside the ionbeam. Also, it is possible that in certain cases when thecoaxial injection of electron and heavy ion beams is

106

104

102

0.03 0.05 0.07 0.09 0.11 0.13

γh/ω

5

106

104

102

4

106

104

102

J 2e,

(cm

2 ⋅keV

⋅c⋅sr

)-1

3

106

104

102

2

106

104

102

1

Fig. 6. Differential electron fluxes with the energy e2e = 101 keV (‘2e’) areplotted in dependence of the normalized growth rate for the resonancedetunings Dx in five frequency ranges: (1) x ’ 0.419x*, (2) 0.457, (3)0.506, (4) 0.618, (5) x ’ 0.708x* (x* � xce/6), f(x)-dashed spline isGNUPLOT exponential function to fit a data-file by the NLLS algorithm,(orbit 202).

N. Baranets et al. / Advances in Space Research 49 (2012) 859–871 867

realized (for example, on equatorial latitudes), a convectiveinstability and ‘evanescent’ waves would be excited. Theamplifying of ‘evanescent’ waves implies that the small-sig-nal disturbances propagate along the system with spatiallygrowing amplitudes (complex k for real x). In most cases,in order to distinguish the absolute and convective instabil-ities developed in considered beam-plasma system, oneshould determine “whether or not Im k(x) has differentsign when the frequency takes on a large negative imagi-nary part” (Briggs, 1964). It take place during asymptoticbehavior of disturbances with an exponentially increasingsinusoidal time dependence.

In Figs. 4 and 5 the data of 201-, 202-experiments arepresented versus the above mentioned frequency ratiop1 = jxm � xbej/xLH. The data shown in Fig. 4 are alsoreported in Baranets et al. (2009) in accordance with theHF electric field fluctuations near the Magion-3. In the fre-quency ranges corresponding to the values of p1 ’ 0.8–1.2all presented data on VLF magnetic and electric field com-ponents blf, elf for the frequencies x/2p = 9.6 kHz and 149,75 Hz, respectively, are enhanced in amplitude (Fig. 4).These VLF waves are registered at point of injection whilethe differential electron fluxes J1e with energy e1e = 245 keVare detected at the distance �110 km from IK-25. Allpresented data look like rapid fluctuations versus theparameter p1 which themselves consists of the fluctuatingfrequencies xbe, xLH. Furthermore, many of the physicalvariables presented on the previous figures are also theshort time-scale fluctuations (in comparison with a sam-pling rate of the data). Thus, the plots in Figs. 4,5 shouldbe considered as results of some virtual experiment withtime-averaged dependencies. The electron flux intensity isslowly decreasing versus the parameter p1 similarly to theamplitude of blf (9.6 kHz) and calculated values of the nor-malized growth rate of electromagnetic instability. Experi-mental results of the lateral and longitudinal quasi-steady(0.1–2 Hz) magnetic and electric field components dEz,dB\ measured at IK-25, respectively, and other data on fastcharged particles (Figs. 4,5) are presented in order to verifythe resonance condition jxm � xbej/xLH � 1 as well as tocompare with the corresponding dependencies of incre-ments ch, ck (Fig. 5). On the one hand, it can be seen thatin the relations Eqs. (11), (13) there is not any singularityrelative to the low-hybrid frequency excitation, on theother hand, a special behavior of measurements near thevalues of p1 � 1 can be found in Fig. 4. If this is true, thenthe conclusion about the reality of strong coupling ofspace-charge beam waves and waves excited due to the beammodulation in the electron gun can be done. The values ofxLH and xpi are very close for the ionospheric conditionsin both cases, but the excitation of low-hybrid frequency ismore preferable for consideration, first of all, because ofthe data distribution versus jxm � xbej/xLH is more clearand logical for decay realization.

It is worth to note also that all presented data do notgive a clear understanding of the triplet wave coupling withfrequencies xm, xbe, xLH. Calculation of the growth rates

ch, ck and charge beam density nbe are produced fordifferent parameters such as an effective modulation depth(R0 not equal to 100% modulation), the electron beamdivergence (D0a0 � 2�–5� vs ape), detuning dx, and other.However, these parameter values have a virtual characterand are important for simulation at every time step beforeto load a complex program. Fluctuations including in Eq.(1) have been maximally smoothed in order to eliminatethe undesirable strong dependencies, because the initialconditions of injection are related to a quiet start and afterthat the beam-plasma instability effects can arise. Never-theless, during 1-s-cycle of electron injection some feedbackin the beam-plasma system must exist. Usage of hollowelectron beam model corresponding to a free electron gyra-tion in high-altitude ionosphere is quite reasonable.

The bursts of fast electron fluxes with the energye1e = 245 keV and slowly falling variations of increment

2.0

4.0

6.0

8.0

γh /ω,×

10−2

106

104

102

100

J 2e,

(cm

2 ⋅ keV

⋅c⋅sr

)-1

106

104

102

100

J 1e

0

100

200

300

400

1.025 1.026 1.027 1.028 1.029 1.030

δB’ ⊥

, nT

kzu/(ω-ωce)

Fig. 7. Normalised growth rates of electromagnetic instability ch/x(h ’ 0.34) (points without lines), fast electron fluxes for two mutuallyperpendicular directions ‘1’ (e1e = 43 keV, ap1 ’ 20�–30�) and ‘2’(e2e = 43 keV, ap2 ’ 60�–70�) as well as disturbances of the lateralmagnetic field component dB\ (at Magion-3) are presented in dependenceof the parameter p3 = kzu/(x � njxcej) (whistler wave excitation condition)for resonance detuning Dx ’ 0.60x* (x* � xce/6), orbit 202. Resultscorrespond to the whistler wave excitation condition ‘ECR’.

868 N. Baranets et al. / Advances in Space Research 49 (2012) 859–871

ch/x along p1-axis are presented for two orbits 201, 202.The similarity of it behavior in the range p1 J 1 was thecause to study new dependencies of the particle fluxes ver-sus the ratio of whistler wave growth rate to the excitedwave frequency p � p2 = ch/x (x ’ xce/6, n = 1,h ’ 0.04).Fig. 6 demonstrates the fast electron fluxes (e2e = 101 keV)as a plot of data points (without lines) registered for per-pendicular direction ‘2e’ relative to z0 at Magion-3 versusthe normalized growth rate of waves excited for the 1stelectron cyclotron resonance (ECR) via normal Dopplereffect. Some of these data points with exponentially grow-ing amplitudes are the evidence of wave-particle interactionfor whistler mode. In Fig. 6 the block data obtained duringthe active experiment on orbit 202 is presented versus theresonance detuning for five frequency ranges. From thesefive data point clusters, one can conclude that the fre-quency range 5 (x ’ 0.708x*, x* � xce/6) is, perhaps, veryclose to whistler mode excitation frequency; the flux data 1(x ’ 0.419x*) demonstrate the absence of any incrementalcharacteristics because of the resonance detuning is verylarge. For relatively dense ionospheric plasma (xpe > xce)the whistler wave excitation due to ECR in accordancewith the Eqs. (11) and (12) is possible for the high-fre-quency range close to xce in dependence of the angle h.The analytical treatment of Eq. (12) to define the whistlermode excitation frequency is very complicated, thus thedata of fast charged particles recorded on Magion-3 subsat-ellite give a good way to verify these processes.

During active experiment with electron beam injectionthe registration of the stimulated fluxes of fast electronsfor both mutually perpendicular directions ‘1e, 2e’ on sub-satellite Magion-3 might be caused due to the excitationprocess of electromagnetic waves and subsequent scatteringduring wave-particle interaction. First mechanism isdefined by the condition of whistler wave excitation byECR on the frequency x = Re x + dx

Rexþ dx ¼ kzuþ njxcej; n ¼ 0;�1; . . . ð14Þ

and second one is defined by the resonance condition ofwhistler wave-fast electron interaction

Rex ¼ kzve cos apj þ mjxcej; m ¼ 0;�1; . . . ð15Þ

where Rex, dx are the frequency of excited waves andresonance detuning, respectively, j = 1,2 and ve mark thenumber of directions ‘1e, 2e’ and velocity of detected elec-trons. If both mechanisms are reliable, the following rela-tion can be considered

kzðu� ve cos apjÞ ¼ dxþ jxcejðm� nÞ: ð16Þ

These wave-particle interactions can be verified during dataprocessing in dependence of the relation p � p3 ffi kzu/(Rex � njxcej) corresponding to the excitation conditionof ECR, or combined relation in accordance with Eq.(16), p � p4 ffi kz(u � vecosapj)/(dx + (m � n)jxcej). Activeexperiment measurements (flux particles, fields) stimulateddue to the resonance conditions (14)–(16) correspond to

p3 � 1, or p4 � 1 obtained with account of the real andsimulation parameter values.

Fig. 7 demonstrates the fast electron fluxes (e1e = e2e

= 43 keV) detected for both directions ‘1e, 2e’ and quasi-lateral magnetic field disturbances dB0? in accordance withthe p3-values (n = 1). More probable that only a certainpart of flux measurements is the result of stimulation dueto the electron injection. A number of measurements(points) presented at this figure is about of 1000 (some ofthem vs p3 are coincident). As concerning the flux datapoints, this presentation is rather a plot of density measure-ments versus p3. For example, the electron fluxes J2e have adensity trend with increasing intensity as the parameter p3

tends to 1, that is when the ECR-resonance conditions (14)are satisfied. If we compare these data with the normalizedgrowth rate for whistler waves (the upper plot of Fig. 7),one can observe that the data of J2e-fluxes correspond lar-gely to the behavior of ch/x. This correspondence isn’t

N. Baranets et al. / Advances in Space Research 49 (2012) 859–871 869

related to the J1e-fluxes registered along z0-axis (close tomagnetic field B00); at the same time a lower limit of J1e-fluxes have a clear tendency to increase as far as the reso-nance mismatch increases. The growth rate calculation isperformed for above-mentioned case 2 of detuning selec-tion when dx � dx0. In this case the growth rate changesare caused not only by the ionospheric parameter varia-tions but also by changes of stream velocity u (not pre-sented here), i.e. by the EG acceleration voltage.Variations of quasi-lateral magnetic field component dB0?are presented in the lower part of Fig. 7. When the reso-nance detuning is a minimal, the amplitudes of dB0?-distur-bances reach to �500 nT what proves the strong relationwith the resonance conditions of whistler wave excitation.In spite of the quasisteady magnetic field disturbances arerelated to the Alfven waves, the excitation of whistlersmaintains a wide range of ULF waves; it was shown inFigs. 4,5.

Fig. 8 demonstrates the anomalous fluxes of fastelectrons in dependence of combined condition Eq. (16)

106

104

102

100

J 1e,

(cm

2 ⋅keV

⋅c⋅sr

)-1

1

106

104

102

J 1e

2

106

104

102

100

J 2e,

(cm

2 ⋅keV

⋅c⋅sr

)-1

3

106

104

102

100

10-2

-1 -0.5 0 0.5 1 1.5 2

J 2e

|kz|(u-vecosαp2)/|ωce|

4

Fig. 8. Fast electron fluxes with the energies e1,2e = 101 keV (panels 1, 4),e1e = 63 keV (2), and e2e = 180 keV (3) are presented in dependence of theparameter p4 ffi jkzj(u � vecosap2)/jxcej (m=0,n=1,dx xce), ve is thevelocity of electrons for energy e2e = 101 keV, orbit 202. Fluxes obtaineddue to the double resonance condition ‘ECR+CR’ correspond to theparameter values p4 � 1 � 1.5.

corresponding to the whistler wave excitation and a subse-quent wave-particles interaction via Cerenkov resonance(‘ECR + CR’, n = 1, m = 0, h ’ 0.4, j = 2). It is worth toemphasise that the real values of p4 might strongly differfrom 1 in dependence of the velocity and pitch-angles ofregistered particles for j = 1,2, that is all variety of theanomalous fluxes may be correctly investigated by thismanner only for some velocity group particles. This meansthat the anomalous electron flux spectra versus p4 can bepresented only by the average virtual spectra. Anomalouselectron fluxes presented in Fig. 8 for p4 ’ 1–1.5 can berelated to the particles really disturbed due to‘ECR + CR’-mechanism. Besides, the behavior of anoma-lous fluxes versus p4 have a regular character and stronglydiffer from previous. The data for p4 � �(0.5–1) and >2can be interpreted as fluxes disturbed due to another inter-action mechanism. Anomalous fluxes registered for theenergy e2e = 101 keV are subject to strong suppression forthe parameter p4 ’ 1–1.5 which was determined for thisparticle energy (panel 4 in Fig. 8), while the fluxes forenergy e1e = 101 keV and j = 1 (plot 1 in Fig. 8) are simul-taneously measured flux particles as well as the particles fore2e = 180 keV (3) and e1e = 63 keV (2). As concerning thedifferent character of stimulated electron fluxes (amplifica-tion and suppression) registered for considered resonancecases, one can assume that these charged particle registra-tion phenomena are the consequence of wave-particleinteraction with right-handed circularly polarized electro-magnetic waves the result of which is, first of all, thepitch-angle scattering. This conclusion is in contrast tothe electron acceleration mechanism proposed in Prechet al. (2002) (E � B-drift across the magnetic field lines)to explain the electron fluxes observed on Magion-3 duringAPEX-experiment. Moreover, it is not possible to evaluatethe electron acceleration in term of short-time interactionwith whistler waves in the frame of this paper. However,as it was noted the only part of the registered electron fluxdata prove the existence of wave-particle resonance interac-tion via proposed mechanism, i.e. it is needed to take intoaccount a wide range of possible origins for charged parti-cle acceleration during 10 keV electron beam injection.

The electron flux spectra are shown in Fig. 9 for the dif-ferent parameter values in the range p4 ’ 1–1.5. The inten-sity of electron flux spectra is strongly increased in theenergy range ee = 20–30 keV for all directions 1, 2, butthe ones are decreased in the middle part of spectrum. Sim-ilar a low-energy electron flux depression was observed inactive space experiment with HF radiation by the dipoleantenna to study a sounder-accelerated particles (SAP)(Baranets et al., 1998). However, the electron flux decreas-ing have been registered during �10 ms when HF emissionwas switch off. Effect have been interpreted as an electronflux scattering in strongly turbulent ionospheric plasmaregions. Investigation of beam-induced effects can be con-tinued with involving a new data of ULF–VLF, HF wavemeasurements registered at the subsatellite, after that, amore conclusions can be made.

Fig. 9. Electron flux spectra recorded in the parameter range p4 ’ 1.15 (a,b), and p4 ’ 1.41 (c,d), dashed curves are the unperturbed flux level J 01e; J 0

2e,orbit 202.

870 N. Baranets et al. / Advances in Space Research 49 (2012) 859–871

We have considered a number of effects connected withthe whistler wave excitation during the electron beam injec-tion in complex beam-plasma system shown in Fig. 1. Forsimplicity, the dispersion equation corresponding to thetransverse-wave instability was restricted to the injectioncase of warm beam electrons in a cold ionospheric plasma.Excitation of backward-propagating whistler waves havebeen determined in term of the absolute instability(real k). In result of these very important approaches, itwas able to estimate the growth rate of whistler wavesamplifying anywhere within the beam-plasma system(backward-wave oscillator). We assumed that this case ofinstability is more possible for the observed ionosphericconditions. However, it does not exclude in part the ampli-fication process due to the convective instability whichcould also contribute the effects to the observed data ofactive experiment.

5. Conclusions

Experimental observations of the anomalous fluxes offast charged particles, disturbances of quasi-steady andULF–VLF magnetic field components, and thermalplasma ion fluxes have been considered using the data of

active space experiment during electron beam injectionthrough an ion beam. These results were obtained fortwo orbits (201, 202) with very similar ionospheric plasmaparameters. Main results of active experiment carried outon mother-daughter satellite system can be formulated asfollows.

First of all, the anomalously large disturbances �400–500 nT of the quasilateral magnetic field component areregistered at Magion-3 subsatellite (at the distance fromthe main satellite d J 100 km) caused by the electronbeam injection (�0.1 A, �10 keV). It is supposed that thenature of these disturbances are related to the conditionsof whistler (Alfven) wave excitation.

During electron beam injection from the main satelliteIK-25, the anomalous fluxes of the fast electrons weredetected by the charged particles spectrometer mountedat the Magion-3 subsatellite. Electron flux perturbationsin the energy range �20–30 keV are caused by the whistlerwaves excited due to the 1st electron cyclotron resonancevia normal Doppler effect; then backward-propagatingwaves can interact with electrons by Cerenkov resonance.The result of these two considered mechanisms is theappearance of the anomalous electron fluxes up to�100 keV.

N. Baranets et al. / Advances in Space Research 49 (2012) 859–871 871

Variations of the thermal plasma ion fluxes weredetected by the ion trap at remote region of injection(Magion-3) in �x0-direction (opposite to a subsatellitevelocity). On the data of experiment in orbit 201 one canconclude that the thermal plasma ions are very sensitiveto ELF–VLF wave excitation during xenon ion beaminjection. This effect is not confirmed by the resultsobtained on 202 orbit. Abrupt change of these ion fluxesis observed when the electron beam modulation frequencysweeps to xm/2p = 62.5 kHz which is close to �xce/2p/6 � 90–100 kHz used in this paper as a whistler excitedwave frequency. Difference of these frequencies leads tothe resonance detuning. We assume, that the thermalplasma ion fluxes are disturbed by the long time-scale wavepacket generated due to a whistler wave nonlinearcoupling.

One of the most important result is related to the ULF–VLF wave effects observed in the frequency range when atriplet wave relation corresponds to jxm � xbej/xLH � 1.These wave phenomena can be interpreted as a parametriccoupling of space-charge beam and undulating waves.When a beat wave excitation occurs near the low-hybridfrequency, the wide band of ULF–VLF frequencies areexcited.

Acknowledgements

The authors would like to thank V.S. Dokukin, andYa.P. Sobolev for information on active APEX regimesand useful discussion of the paper results. One of us(K.K.) wishes to acknowledge VEGA grant agency project2/0081/10 for support.

References

Baranets, N.V., Sobolev, Ya.P., Ciobanu, M., et al. Plasma Phys. Rep. 33,995–1013, 2007.

Kitsenko, A.B., Stepanov, K.N. Ukr. Fiz. Zh. 6, 297–303, 1961.Baranets, N.V., Sobolev, Ya.P., Ruzhin, Yu.Ya., et al. J. Plasma Fusion

Res. SERIES 8, 251–256, 2009.Maehlum, B.N., Maseide, K., Aarsnes, K., et al. Planet. Space Sci. 28,

269–279, 1980.Grandal, B. Artificial Particle Beams in Space Plasma Studies. Plenum,

New York, 1982.Marshall, J.A., Lin, C.S., Burch, J.L., et al. J. Space Rockets. 25, 361–367,

1988.Oraevskiy, V.N., Sobolev, Ya.P., Zhuzgov, L.N., et al. Plasma Phys. Rep.

17, 699–705, 2001.Kaptsov, N.A. (Ed.), Radiophysical Electronics (Mir, Moscow, 1992) (In

Russian).Oraevskiy, V.N., Ruzhin, Yu.Ya. (Eds.), Project APEX. Scientific

Purposes, Simulation, Technology and Equipment of Experiment(Nauka, Moscow, 1992) (In Russian).

Banks, P.M., Raitt, W.J. J. Geophys. Res. 93, 5811, 1989.Mishin, E.V., Ruzhin, Yu.Ya., Telegin, V.A. (Eds.), Interaction of

Electron Fluxes with the Ionospheric Plasma (Gidrometeoizdat,1989) (In Russian).

Mikhailovskiy, A.B. (Ed.), Theory of Plasma Instabilities V.1. (Atomiz-dat, 1975) (In Russian).

Roeder, J.L., Sheldon, W.R., Benbrook, J.R., et al. Ann. Geophys. 36,401, 1980.

Mett, R.R., Tataronis, J.A. Phys. Rev. Lett. 63, 1380, 1989.Moffatt, H.K. Magnetic Field Generation in Electrically Conducting

Fluids. Cambridge Univ. Press, 1978.Briggs, R.J. Electron-stream Interaction with Plasmas. The M.I.T. Press,

Cambridge, Massachusetts, 1964.Prech, L., Nemecek, Z., Safrankova, J., Omar, A. Ann. Geophys. 20,

1529–1538, 2002.Baranets, N.V., Galperin, Yu.I., Erokhin, N.S., et al. Adv. Space Res. 21,

709–712, 1998.